Criteria abstracted from The Users' Guide to Medical Literature, from the Health Information Research Unit and Clinical Epidemiology and Biostatistics, McMaster University

Highlighted lines and questions below provide links to the pertinent description of criteria in The EBM User's Guide, now available at the Canadian Centres for Health Evidence

Article Reviewed:

Please visit the new Evidence Based Journal Club Reviews

Risk factors for long intensive care unit stay after cardiopulmonary bypass in children.

Brown KL, Ridout DA, Goldman AP, Hoskote A, Penny DJ.

Crit Care Med. 2003;31(1):28-33. [abstract]

Reviewed by Al Torres, MD, MS, University of Illinois College of Medicine at Peoria

Review posted April 29, 2004

I. What is being studied?

Study objective:

• To determine pre-, intra-, and post-operative risk factors of children who experience longer ICU stays after open heart surgery.
• To identify factors associated with longer ICU stays that are potential targets for quality improvement.

Study design

A retrospective review of the pre-, intra-, and postoperative factors was performed for 355 children undergoing open heart surgery from April 1, 1999 to March 31, 2000 (1 year) at Great Ormond Street Hospital for Sick Children.

II. Are the results of the study valid?

Note: These questions follow from Randolph AG et al. Understanding articles describing clinical prediction tools. Crit Care Med 1998;26:1603-1612. [abstract]

1. Was a representative group of patients completely followed up? Was follow-up sufficiently long and complete?

Unclear. The investigators do not provide a detailed description of the patients or their operative procedures which would permit a comparison to one's own patient population. The average age of a cardiac ICU population has grown younger (weeks rather than months old), and primary repairs are being selectively performed over palliative procedures. Follow-up was complete for the 342 patients included, i.e., to cardiac ICU discharge. The investigators do not describe their ICU discharge policy such as whether or not patients are discharged to another hospital unit after extubation and/or discontinuation of inotropic support or they discharged directly from the ICU to home.

2. Were all potential predictors included?

No. The investigators included clinical variables from the pre-, intra-, and postoperative period that would influence outcome (see table below). However, the investigators did not include or describe the process of care for their particular ICU. For example, the presence of an intensivist or physician extender 24 hours a day has been reported to facilitate extubation and minimize morbidity which influence length of stay in the ICU. The investigators created and included a scale of operative procedure complexity (see the appendix) which assigned a higher numeric category for increasingly complex procedures (e.g., Category 1 for ASD repair, Category 6 for Stage I Norwood).

Preoperative Factors	Intraoperative Factors	Postoperative Factors
Presence of 2 medical problems	cross-clamp time	delayed sternal closure
Down's syndrome	bypass time	reoperation within the same PICU admission
Mechanical ventilation	circulatory arrest	postoperative complications
Resuscitation < 24 hours pre-op	operative complexity	pulmonary hypertension

3. Did the investigators test the independent contribution of each predictor variable?

Yes. The investigators defined a significant association of almost all potential risk factors with length of stay using univariate analysis and then incorporated these factors into a multivariate analysis (e.g., multiple regression analysis).

4. Were outcome variables clearly and objectively defined?

Yes. The investigators defined length of stay as the cardiac ICU discharge date minus the date of admission for the purposes of, or immediately after open heart surgery. They defined those children who fell in the top 5% for length of stay as the long-stay patients. The investigators did not provide any criteria used to define ICU discharge readiness.

III. What are the results?

1. What is(are) the prediction tool(s)?

The investigators performed an overall analysis of the pre-, intra-, and postoperative factors and identified the most important factors independently associated with length of stay to be: a requirement of preoperative ventilation, neonatal status, preoperative medical problems, cardiopulmonary bypass time, operative complexity category, and a postoperative complication score (which was not well defined). They did not go the next step and develop a true model or prediction tool to identify risk.

2. How well does the model categorize patients into different levels of risk?

The investigators did not build a model per se that allows us to assign a risk category to an individual patient (e.g., a total score of 12 puts the patient at a 40% risk of long stay). Part of the problem is that the outcome measure is a continuous, non-normally distributed variable - the length of stay. This question is more geared to outcomes that are binary, e.g., mortality.

3. How confident are you in the estimates of the risk?

Estimates of risk of prolonged length of stay were not performed. The investigators only reported the 95% CI for the incident rate ratio, i.e., the ratio of the average length of stay for those children with the particular risk factor present compared with the average length of stay for those children without that particular risk factor (see table).

Final model	Adjusted IRR	95%CI for IRR	p value
Preoperative ventilation	1.50	1.15, 1.97	< .003
Neonate	1.38	1.09, 1.73	.007
2 medical problems	2.61	1.87, 3.65	< .001
CPB time, 30 min	1.08	1.02, 1.13	.004
Complications score	1.66	1.53, 1.81	< .001
Operative class, 4, 5, 6, vs. 1, 2, 3	1.28	1.06, 1.54	.009

These confidence intervals were, in fact, rather narrow. For example, subjects with preoperative ventilation have a length of stay 1.15 to 1.97 times that of subjects without preoperative ventilation.

IV. Will the results help me in caring for my patients?

1. Does the tool maintain its prediction power in a new sample of patients?

Unknown. The investigators did not create a prediction tool to discriminate patients at risk for length of cardiac ICU stay > 14 days from those at lower risk. It would be helpful to see if the factors independently associated with prolonged length of stay remain associated in a separate population from another cardiac ICU.

2. Are your patients similar to those patients used in deriving and validating the tool(s)?

Maybe. As stated in II.1., the investigators do not provide a detailed description of the patients or their operative procedures to permit a comparison. The results may be different in a cardiac ICU with a much smaller population with different resources and processes of care.

3. Does the tool improve your clinical decisions?

Yes. The investigators wisely point out that when ICU bed availability is an issue, children with complex medical conditions and/or complex operative procedure should be electively scheduled for surgery in series rather in parallel when possible. The investigators also described a "domino effect" in length of stay when several important postoperative complications occur together in the same patient. For example, patients who suffer a cardiopulmonary arrest will also likely suffer from renal failure, an arrhythmia, and pulmonary hypertension which were all strongly associated with prolonged length of stay.

4. Are the results useful for reassuring or counseling patients?

Maybe. If the identified pre-operative risk factors are confirmed in a separate population, parents of children with multiple risk factors for prolonged ICU stay can be informed preoperatively that there is an increased likelihood that their child may be in the cardiac ICU for a prolonged period of time. Likewise, the occurrence of these intra- and post-operative risk factors could be used to prepare families and predict resource use.

 

Comments

Back to the EB Journal Club Index